The Seed Bank at Summer's End: A Reflection on Generative Preservation
There’s a particular quality to the light in late summer, a golden slant that seems to hold the season’s fullness while whispering of its imminent end. In my local park, the sunflowers are bending their heavy heads, their work nearly done. It’s during these weeks that I find myself thinking less about preserving what is, and more about ensuring what can be. It’s the time for harvesting seeds.
This annual ritual has led me to a parallel thought about our digital world, particularly what we call preservation. So much of our effort, rightly so, is like that of a photographer capturing the perfect bloom—a static snapshot, a moment frozen in a PDF, a website perfectly preserved in amber. We are brilliant at building archives that are, in essence, vast and beautifully organized herbaria. But a herbarium specimen, while invaluable for study, cannot grow. It cannot adapt to a new climate or cross-pollinate with a neighboring plant. Its life is over; its information remains.
What, then, is the digital equivalent of a seed bank? I ponder this while emptying dried seeds from a papery sunflower head into a labelled envelope. It’s a form of preservation, yes, but a fundamentally different one. It’s not preservation for reference, but preservation for regeneration. The value isn’t in the seed as an object, but in the latent, living potential inside it—the dormant code that can, given the right conditions, sprout into a whole new entity.
In our digital vernacular, this is the preservation of source data, not just the rendered page. It’s the open dataset behind the now-defunct interactive map, the raw text and markup of a blog before it was squashed into a proprietary social media feed, the original, uncompressed creative file. This is the ‘seed’ data: messy, often requiring context and care to ‘germinate,’ but containing the full genetic blueprint of the original creation. It allows for future reinterpretation, for new visualizations, for adaptation to platforms we haven’t even dreamed of yet.
The challenge, of course, is that seeds require more active care than a pinned butterfly. They need to be kept viable, their formats documented, their dependencies understood. A JPEG from 1995 might be perfectly viewable today, but a raw dataset from a custom database in 2010 might be functionally extinct if the schema is lost. This is the quiet, unglamorous work of the digital seed banker: ensuring not just that the bits survive, but that the knowledge of how to make them live again survives with them.
As I seal the envelope and note the date and variety, I think about the coming winter. The garden will die back, but the potential for next year’s bounty is tucked safely away. Our digital autumn is constant; things are always ending. By thinking like a seed saver, we can shift our preservation efforts from simply building museums of the past to stocking the larders of the future. We can ensure that the data we save isn’t just a memorial, but a promise of life to come.
Notes & further reading
A few pages I came back to while writing this: